Electric hoist fault prediction method and system based on double-track interlocking and bidirectional causal verification

Abstract

The present application relates to the technical field of electric hoist fault judgment, in particular to an electric hoist fault prediction method and system based on double-track interlocking and bidirectional causal verification. By constructing the working condition discrete state track and the frequency spectrum continuous state track, the interlocking check is performed to extract effective fault features. The feature sensitivity is sorted by using the inner ring virtual variance disturbance, and is compared with the outer ring historical weight distribution to realize the adaptive reconstruction of the optimal feature weight. At the same time, the forward physical deduction and the reverse historical case retrieval are carried out, and the high confidence fault risk value is output through the dual consistency check of the shape and the rate. The dynamic probability envelope line is constructed, the grading early warning is realized by combining the critical state oscillation monitoring, the model closed loop updating and the parameter self-optimization are completed through the field verification data, and the whole realizes the adaptive, high confidence and self-evolution of the prediction process.

Images